The World Bank

Enterprise Size, Financing PatternsEnterprise Size, Financing Patternsand Credit Constraintsand Credit Constraints

in Brazilin Brazil

Analysis of Data from the Investment Climate Assessment Survey

Anjali Kumar and Manuela Francisco

Enterprise Size, Financing Patterns and Credit Constraints in Brazil

2The World Bank

INTRODUCTIONINTRODUCTION

If access to external financing is based on current performance, or expected future performance, size per se should not have an impact on access to external finance. Yet in many countries it is perceived that small firms face particular disadvantages in the credit market.This paper examines the extent to which firm size affects financing patterns and restricts access to finance in one country, Brazil, based on an Investment Climate Survey of 1642 firms constructed in 2003, which includes firms in thirteen Brazilian states (out of 27) and nine industrial groups.

3The World Bank

Key Questions AddressedKey Questions Addressed

Do small firms financing patterns differ from large firms, and do small firms have less access to credit and face more credit constraints than larger firms?What is the importance of firm size, compared to performance or other factors, in assessing access to credit and credit constraints?Are credit provision criteria different for fixed capital and for working capital (long vs. short-term loans)?Does bank ownership (public, private or foreign) impact differentially upon credit provision across firm sizes?What is the role of credit constraints relative to other constraints, in relation to firm size?

4The World Bank

Factors Affecting Access to Credit Factors Affecting Access to Credit –– a Basic Approacha Basic ApproachTraditional theory suggests that in well-functioning credit markets, lenders would base their decisions on the overall financial soundness of firms and on expected performance and returns, rather than upon firm size.

Yet, empirical studies have also found that smaller and younger firms are more credit constrained than larger and long established firms.

At the cross country level, studies have looked primarily at differentials in debt equity ratios, and results suggest that size does affect financing patterns.

5The World Bank

How Could Firm Size Affect Access to CreditHow Could Firm Size Affect Access to Credit

Greater informational opacity (lack of publicly known contracts, poor credibility of their financial statements) are associated to greater constraints for small firmsTo the extent that the adverse effects of information asymmetry may be reduced by the provision of collateral, it is argued that smaller firms face greater difficulties. Small firms are associated greater real risk compared to large firms (higher failure rate) Such differences are translated into higher bank transaction cost of lending to small firms (search, information, evaluation, monitoring), and higher risk.

6The World Bank

Other Factors Affecting Access to CreditOther Factors Affecting Access to Credit

Industry effects

Regional effects

Ownership effects

Managerial education effects

The lender effect (Bank relationships, Bank ownership)

Other factors (Competitiveness, Credibility, Collateral, and Capacity for Innovation)

7The World Bank

The Dataset: Characteristics of Sample FirmsThe Dataset: Characteristics of Sample Firms

No. firms No. Firms No. Firms No. Firms Sales No. Firms(%) (%) (%) (%) Growth (%)24 127 1549 Post 331 1042(1.5) (7.7) (94.4) Graduate (20.2) (64.6)238 106 86 500 390(14.5) (6.5) (5.2) (30.5) (24.2)121 442 7 Incomplete 249 182(7.4) (26.9) (0.4) University (15.2) (11.3)713 173 Vocational 185(43.4) (10.5) Training (11.3)546 84 158(33.2) (5.1) (9.6)

183 62(11.2) (3.8)79 95(4.8) (5.8)130 Incomplete 60(7.9) Primary

School(3.7)

315(19.2)

Furniture

Auto-parts

Primary School

Electronics

Incomplete Sec. School

Machinery

South Chemicals Secondary School

Southeast Shoes &Leather

Center-West Garments State Unchanged

Increased

Northeast Textiles Private Foreign

Graduate Decreased

Manager’s Education

North Food Processing

Private Domestic

Region Industry Ownership

8The World Bank

The Dataset: Classifications of Firm SizeThe Dataset: Classifications of Firm SizeAs defined by Ministry of Industrial

Development and External Trade and adopted by IBGE and SEBRAE.

An alternative classification used by Brazil’s development

Bank (the BNDES)

In addition, classification of firms by size deciles and quintiles was also investigated.

Number of employees

Sales

(No.s) (R$ 000 per year)

Micro 0 to 19 330 20 < 1,200 736 46

Small 20 to 99 861 52≥1,200 & <10,500 468 30

Medium 100 to 499 376 23≥10,500 & <60,000 268 17

Large More than 500 75 5 ≥60,000 170 7500 -999 53

1000-1999 122000-4999 7

>5000 3Total 1642 100 1642 100

Number of firms

% Number of firms

%

The World Bank

II. FIRM SIZE, FINANCING, ACCESS II. FIRM SIZE, FINANCING, ACCESS TO CREDIT AND CREDIT TO CREDIT AND CREDIT

CONSTRAINTSCONSTRAINTS

10Enterprise Size, Financing Patterns and Credit Constraints in Brazil

10The World Bank

Firm Size and Financing PatternsFirm Size and Financing Patterns

Micro Small Medium Large Micro Small Medium LargeNo. of employees 0-19 20-99 100-499 >500 0-19 20-99 100-499 >500Internal funds 44.2 43.3 44.8 41.2 58.7** 57.8** 54.0** 41.0**

Bank finance1

Foreign 0.8*** 0.9*** 1.7*** 4.9*** 0.0*** 0.8*** 2.6*** 3.2*** Local private 10.8 12.7 12.6 8.5 5.7 6.9 5.4 1.4

Local public2 11.9* 15.2* 17.6* 25.2* 10.4*** 14.1*** 19.1*** 34.5*** Of which government fundsTrade credit 14.2 16.3 13.7 14.2 11.9* 8.6* 6.6* 9.2*Leasing 0.5 0.9 0.8 0.3 2.2 3.1 3.5 5.0Informal sources 10.5*** 5.5*** 1.8*** 0.2*** 4.4*** 2.4*** 0.4*** 0.0***Equity finance 2.7 2.7 4.7 1.8 3.5 3.8 6.0 4.0Credit card finance 0.8 1.0 0.3 0.0 0.5 0.2 0.2 0.0Others 3.6 1.5 2.0 3.7 2.7 2.3 2.2 1.7Total (%) 100 100 100 100 100 100 100 100Total no. of firms 328 860 373 72 247 716 324 64

25.3***

1 This disaggregation does not derive directly from the questionnaire. Local commercial bank finance is disaggregated into local private and local public finance according to the main bank the firm does business with.2 Government funds are included in the local public bank finance category. Statistical significance: * significant at 10%, ** significant at 5%, and *** significant at 1%.

Working Capital New Investments

0.8*** 1.9*** 2.9*** 6.0*** 4.5*** 6.5*** 12.5***

Looking at financing patterns across firms of different size, (i) differentials by size may be more pronounced for fixed capital than for working capital; (ii) a larger percentage of firms among medium and large firms have overdrafts or line of credit, compared to micro and small firms; (iii) separating

banks by ownership, public banks are more significant providers of capital for larger firms.

11The World Bank

Access to Credit and Credit ConstraintsAccess to Credit and Credit ConstraintsSample FrequenciesSample FrequenciesFirms with access to credit are defined as those that express a demand for credit, apply for a bank loan and receive it. Constrained firms are those that express a demand for a bank loan but either (i) apply for a bank loan and are rejected, or (ii) do not apply. This is access to credit in a narrow sense.About 54% of large firms that did not apply for credit reported that they did not need a loan, compared to 39% of micro firms. About 61% of micro firms that did not apply for a bank loan reported other reasons (such as application procedures, collateral requirements, interest rates, or expectations of being rejected) compared to 46% among large firms.Only 2.7% of large firms did not have a loan because their application was rejected, compared to 9.4% for micro firms. About 38% of micro firms did not apply for bank loans (even though they needed one) because of other reasons cited above. For large firms that percentage corresponds to 18%.

12The World Bank

Relative Importance of Factors Affecting Access to Relative Importance of Factors Affecting Access to Credit: A Simple ModelCredit: A Simple Model

To test whether size, performance, industry, region and education explain the probability of having a loan, we estimate a maximum likelihood probit model incorporating these variables, and estimate the marginal effects of these variables on access to credit as defined above.The results:

Firm size dominates all other effects.An estimation of the marginal effects shows that being a small, medium or large firms increase the probability of having a loan compared to micro firms by 9, 22, and 34 percentage points respectively.

13The World Bank

Relative Importance of Factors Affecting Relative Importance of Factors Affecting Access to Credit: An Alternative ModelAccess to Credit: An Alternative ModelIn order to test the robustness of the results, we use a two step maximum likelihood probit with sample selection.

The probability of having demand for a bank loan is estimated in the first stage.

In a second stage, the model estimates access to credit. Having access to credit is defined by having a bank loan.

14The World Bank

Relative Importance of Factors Affecting Access to Relative Importance of Factors Affecting Access to Credit: An Alternative ModelCredit: An Alternative Model

The Specification:Demand for credit==α+β firms’characteristics + δ banks’ evaluation of firms’ risk + ε

Firm’s characteristics: firm size, region, industry, ownership, managers’ education, capacity utilization, age, exporter, corporate status, group membership, innovative capacity (percentage of workers that use a computer regularly).

Firm’s willingness to invest proxies: whether a bank has an overdraft or line of credit, the percentage of inputs bought on credit, and cited macroeconomic obstacles to growth (economic uncertainty, macroeconomic instability, and cost of credit).

15The World Bank

Relative Importance of Factors Affecting Access to Relative Importance of Factors Affecting Access to Credit: An Alternative ModelCredit: An Alternative Model

The Specification:Access to credit=α+β firms’characteristics + δ firm’s willingness to invest + ε

Firm’s characteristics: firm size, region, industry, ownership, managers’ education, capacity utilization, age, exporter, corporate status, group membership, innovative capacity (percentage of workers that use a computer regularly).

Firm’s risk: performance variables (turnover, sales growth, leverage), information transparency (external auditor), the nature of the banking relationship (unique or not), and whether the firm has an overdraft and collateral or not

16The World Bank

Relative Importance of Factors Affecting Relative Importance of Factors Affecting Access to Credit: An Alternative ModelAccess to Credit: An Alternative ModelThe most relevant results:

Being a large or medium firm increases the probability of having a loan by 25 and 15 percentage points compared to micro firms, respectively. Having an overdraft has a positive impact on the probability of having a loan (by 16 percentage points).Having a unique bank relationship decreases the probability of having a loan, by 11 percentage points.An increase in the workforce education increases the likelihood of having a loan by 3 percentage points.When splitting the sample according to loan duration, none of the variables seems to explain access to credit to short term loans.

The World Bank

III. FINANCIAL INSTITUTION III. FINANCIAL INSTITUTION OWNERSHIP AND ACCESS TO OWNERSHIP AND ACCESS TO

CREDITCREDIT

18The World Bank

Firms, Access to Credit and Credit ConstraintsFirms, Access to Credit and Credit ConstraintsBreakdown per Type of BankBreakdown per Type of Bank

Type of Institution No. of Firms %Domestic Private Banks 687 42.3Foreign Private Banks 207 12.7Public Banks 725 45.0Total 1626 100

Private Domestic Bank

Private Foreign Bank Public Bank

Have a loan (%) 42.4*** 44.9*** 53.4***Loan application rejected 20.8** 14.3** 12.6**Constrained 55.8*** 53.1*** 43.8***Have overdraft 70.1*** 75.8*** 79.6***Required collateral 67.3 65.2 67.4

Statistical significance: * significant at 10%, ** significant at 5%, and *** significant at 1%.

Banco do Brasil, a public domestic bank, is the principal bank for 593 firms, or 36% of the total sample. It is also the most important financial institution for small firms, though micro firms appear to engage most importantly with the Caixa Economica Federal, the second largest bank, also publicly owned. In contrast to Banco do Brasil, CaixaEconomica Federal’s clients include few mid sized firms and no large firms. The second most important bank for firm of all sizes is Bradesco, a privately owned domestic bank. Its importance as the main bank does not vary across firm size.

19The World Bank

Testing whether Access to CreditTesting whether Access to CreditVaries According to Bank OwnershipVaries According to Bank OwnershipA. Split the sample according to:(i) bank ownership (public banks versus private banks);

Among the firms that work primarily with public banks, large firms are the most likely to have a bank loan. We do not find a similar result for large firms that work primarily with private banks. Sales growth, innovation capacity, and overdraft are positively associated with the likelihood of having a loan if the bank they work with is a private bank. Among the firms that work primarily with a private bank, new firms are less likely to have a bank loan.

(ii) interactive dummy: firm size*bank ownershipSmall firms that work primarily with public banks are not more likely to have a bank loan than small firms that work with private banks.

The World Bank

IV. FINANCIAL ACCESS AS AN IV. FINANCIAL ACCESS AS AN OBSTACLE TO GROWTH COMPARED OBSTACLE TO GROWTH COMPARED

TO OTHER VARIABLES TO OTHER VARIABLES

21The World Bank

Financial Access as an Obstacle to GrowthFinancial Access as an Obstacle to GrowthCompared to Other VariablesCompared to Other Variables

The cost of financing is reported to be the main obstacle to growth (57% of firms). Access to financing is ranked seventh (34.5%) after cost of financing, tax rates, corruption, economic and regulatory policy uncertainty, and macroeconomic instability.Looking at the differential impact of obstacles across firm size it appears that access to financing and costs of financing are smaller obstacles to growth for larger firms relative to other sizes. Furthermore, larger firms are less likely to rate tax rates and corruption as very high obstacles to growth.

The World Bank

V. CONCLUSION V. CONCLUSION

The World Bank

ConclusionConclusionSources of finance vary by firm size, and moreover, size may affect access to investment financing more strongly than to working capital financing.Size is an important determinant of credit access and credit constraints. Moreover, size appears to have a much more significant effect on determining access to credit than performance-related variables. Also, there is an effective quantitative limit in the allocation of credit to smaller sized borrowers.Among firms that work primarily with public banks large firms are the most likely to have a bank loan. Moreover, small firms that work with public banks are not more likely to have a bank loan than small firms that work with private banks.Other reasons such as taxation and corruption are also important as obstacles to growth for all firms.

The World Bank

THANK YOU!THANK YOU!

The World Bank

APPENDICESAPPENDICES

The World Bank

ReferencesReferences

Firm size effect on access to creditBinks and Ennew (1996), Berger and Udell (1993, 1994, 1998), Saito and Villanueva (1981), Stiglitz and Weiss (1981), Lung and Wright (1999)Angelini et al. (1998), Berger and Udell (1994), Bester (1987), Binksand Ennew (1996), Avery et al. (1998), Binks and Ennew (1996)Lund and Wright (1999), Gertler and Gilchrist (1994), Chittenden et al. (1993), Hughes and Storey (1994), Klapper et al. (2002), Brewer et al. (1996), Saito and Villanueva (1981)Saito and Villanueva (1981)

Assessing factors affecting access to creditBigsten et al. (2003), Levenson and Willard (2000), Levy (1993)Demirguc-Kunt and Maksimovic (1999), Beck, Demirguc-Kunt and Maksimovic (2002a,b), Demirguc-Kunt and Maksimovic (2001)

The World Bank

ReferencesReferencesOther Factors Affecting Access to Credit

Industry effect: Banks may favor firms of specific industries as clients, lending more to ‘growth’; some industries are more likely to depend on external financing than others - Rajan and Zingales(1998), Bigsten et al. (2002). Regional effect: From differentials in bank density across regions, which themselves may reflect differentials in income and levels of economic activity - Beck (2002), Kumar et. al (2004).Ownership effect: Foreign firms may have more access to credit and less credit constraints than domestic private firms. State firms may have more credit access (especially from public banks) relative to private firms; but if access to credit depends on performance, state firms should be more credit constrained than private firms - Willmore (1986), Majumbar (1998), Vinning and Boardman (1990). Managerial education effect: This has not been addressed in previous empirical studies. Various authors have raised its importance - Jensen and McGuckin (1997), Kumar (2004).The lender effect: Closer banking relationships could reduce transaction costs that emanate from information asymmetries - Binks and Ennew (1996, 1997), Peterson and Rajan (1994), Chakravartyand Scott (1999), Cole (1988), Peterson and Rajan (1994), Ferri and Messori (2000). Another factor is the ownership of the lending financial institution, as foreign banks may provide more credit to large corporate firms and public banks to small firms by mandate - Berger et al. (2001), Clarke et al. (2001, 2002)Other: Competitiveness reflected in age, (Frazer, 2004), whether firms are exporters or not, firms’transparency and credibility (Binks and Ennew, 1996), innovation and technological change (Solow, 1957, Long et al. 2003, Griliches, 1998, Barro and Sala-i-Martin, 1995, Lucas, 1988, Romer, 1990. Laursen et al. ,1999.

Relative Importance of Factors Affecting Access to Credit: A SimRelative Importance of Factors Affecting Access to Credit: A Simple Modelple Model

Having a loan Size Firm Ownership Small firms 0.090** State firms 0.212 (2.20) (0.91) Medium firms 0.222*** Foreign firms -0.031 (4.56) (0.38) Large firms 0.338*** Education (4.28) University degree 0.005 Industry (0.10) Food processing 0.110* Incomplete university -0.056 (1.75) (1.03) Textile 0.098 Vocational training after secondary school 0.003 (1.20) (0.05) Shoes and Leather products 0.096 Secondary school -0.103 (1.28) (1.62) Chemicals 0.034 Incomplete secondary school -0.053 (0.37) (0.61) Machinery 0.075 Primary School -0.002 (1.03) (0.03) Electronics 0.099 Incomplete primary school 0.010 (1.03) (0.12) Auto-parts 0.016 Performance (0.21) Sales growth1 -0.030 Furniture 0.053 (0.96) (0.82) Observations 1116 Region South 0.068* (1.89) Center-West -0.006 (0.09) North -0.186 (1.26) Northeast -0.005 (0.11)

Relative Importance of Factors Affecting Access to Credit: An AlRelative Importance of Factors Affecting Access to Credit: An Alternative Modelternative Model Having a loan1

Size Small 0.034 (0.63) Medium 0.154** (2.33) Large 0.248*** (2.42)

Performance Turnover (sales/assets) -0.000 (0.88) Leverage 0.000 (0.20) Sales growth 0.001 (1.27) Firm characteristics Exporter 0.002 (0.05) SA 0.034 (0.33) Group 0.038 (0.49) Capacity utilization 0.001 (0.86) New firm -0.077 (1.37) % workforce that use computers 0.003** (2.44)

External auditor 0.050 (0.90) Collateral 0.014 (0.32) Relation with banks Overdraft 0.158*** (3.40) Bank unique relationship -0.111** (2.02) Other Control variables: Industry Region Firm ownership Education

Yes Yes Yes Yes

Observations 1088 Wald chi2 74.55

Access to Credit and Credit ConstraintsAccess to Credit and Credit ConstraintsSample FrequenciesSample Frequencies

Micro Small Medium Large

Have a bank loan (% of total no. of firms) 27.1 31.9 43.9 58.9 Do not have a loan Rejected (% of do not have a loan) 12.9 10.2 8.6 6.7 Did not apply (% of do not have a loan) 87.1 89.8 91.4 93.3 Total (% of do not have a loan) 100 100 100 100 Did not apply No need (% of did not apply) 39.2 44.5 51 53.6 Other reasons2 (% of did not apply) 60.8 55.5 49 46.4 Total (% of did not apply) 100 100 100 100 Total of firms constrained (% of total no. firms) 47.7 40.7 29.7 20.5 Application was rejected (% of total no. firms) 9.4 7.0 4.8 2.7 Did not apply (% of total no. firms) 38.3 33.7 24.9 17.8 Reasons for Not Applying for a Bank Loan Do not need loans 39.4 44.7 51.3 53.6 Applications procedures 13.5 9.4 7.3 0.0 Collateral requirements 7.7 9.2 4.2 3.6 Interest rates are too high 36.5 33.2 33.0 32.1 Corruption in the allocation of bank credit 1.0 0.6 0.0 3.6 Did not expected to be approved 1.4 0.8 0.5 0.0 Others 0.5 2.3 3.7 7.1 Total (%) 100 100 100 100 Reasons for Bank Loan Rejection Lack of collateral 41.9 41.7 22.2 50.0 Incompleteness of the application 19.4 16.7 11.1 0.0 Lack of feasibility of the project 3.2 10.0 22.2 0.0 Poor credit history 25.8 21.7 38.9 50.0 Others 9.7 10.0 5.6 0.0 Total (%) 100 100 100 100